Navigation App Redesigned: L1→L5 Autonomous UX and Agents

A live demo of a Toronto navigation app showing L1‑L5 autonomy levels, using a verbs‑first interface with intent bar, predictive chips, and autonomous rerouting.

Overview

I’ll run a local build of a Toronto-focused, AI-assisted navigation app and walk through the code that implements five levels of autonomy (L1–L5) using a verbs-first UX.

Demo flow:
Current Maps UI vs. agent-driven future (verbs, not forms)
L2 Supervised: Mock incident → side-by-side reroute modal (p50/p90 ETA, reliability %, chaos) → Approve with haptics → “Why” provenance chips
L3 Conditional: Auto-reroute + 10s Undo Snackbar (graceful action stacking)
Verbs-First Input: Intent Bar (“Arrive by 8:45”), Predictive Chips, ⌘K Command Palette, Leave-By Dial with live on-time probability
L4 Geofenced: Downtown zone activates autonomy banner + Operations Timeline
L5 Full Autonomy: Morning Digest observer mode with Pause AI toggle

Code will be running live. No pitch/slides just app + demo.

Tech stack